Standardise plot.pacea_st() and plot.pacea_oi() with respect to colour bars

[andrew-edwards]

Needed to check #60.

One function seems to use the filtered data (what you've chosen to plot) to set the colour bar, and one uses the full data set. Want to do the same because when mapping OISST onto grid26 (#60), the OISST data become a pacea_st object that uses plot.pacea_st().

Also nice to have customisable options. So three options for both functions would be:

• "global" (default), use the range of the input data, regardless of what is being filtered to plot
• "filtered" - just the range of what is being plotted (i.e. months and years, spatial will be done before function I think)
• user-defined, so could be a vector of the range.
  Then will be very customisable.

Steps are:

• add options to plot.pacea_st() and plot.pacea_oi()
• do some plots in oisst_to_grid26.Rmd to check the grid26 mapping worked well
• redo oisst and bccm vignettes locally to see that they look somewhat similar
• update help files
• adapt vignettes, maybe just adding in some comments about the options
• add to main README and to NEWS